This invention relates to a spatial light modulation device used for an image processing operation in a tomographic apparatus, and to an image reconstructing apparatus using the spatial light modulation device.
Heretofore, in the field of image processing, images are, in general, digitally processed by using a computer system. One application of such an image processing technique is an image reconstructing apparatus as shown in FIG. 6, which is used in a tomographic apparatus.
In the image reconstructing apparatus as shown in FIG. 6, a ray source 80 applies, for instance, X-rays to an object 81 under examination, and a measuring instrument 82 detects a dose of X-rays passed through the object. The dose of X-rays thus detected is a projection datum produced by adding internal information of the object in the direction of irradiation and projecting the added result. The projection datum is converted into a digital value by an A/D (analog-to-digital) converter 83, which is provided to a comparator 84. In the image reconstructing apparatus, an image memory 85 is used to store a reconstructed image. A reconstructed image has picture elements the number of which corresponds to the number of internal information of the object. The data f(i,j) of the picture elements are summed in the direction corresponding to the direction of irradiation of the object 81, and provided to the comparator 84. At the beginning, the data f(i,j) of the picture elements have suitable initial values. In the comparator 84, the projection datum is compared with the sum of the data f(i,j) of the picture elements taken in the direction of irradiation to obtain the difference value therebetween. The difference value is provided to a computer system 86. In the computer system 86, the difference values are subjected to a prescribed digital operation to produce correction data. The correction data are laid over the data f(i,j) of picture elements to correct the data f(i,j) so that the difference between each internal information of the object 81 and the corresponding datum f(i,j) of picture element may be decreased.
Thereafter, either the ray source 80 or the object 81 is rotated by a predetermined angle .omega., and according to the projection data which are obtained with the changed direction of irradiation and the reconstructed image, the above-described operations are carried out again. In this manner, the reconstructed image is sequentially renewed; that is, the image processing operations are carried out until the difference between the reconstructed image and the internal information of the object is less than a predetermined threshold difference.
The above-described image reconstructing method, being disclosed in the literature "Iwanami Lectures, Information Science--21, (Pattern Recognition and Figure Processing)", pp. 192 to 195 published on Mar. 10, 1983, is well known as an iterative approximation method in the field of image processing.
In the image reconstructing apparatus according to the conventional iterative approximation method, the image processing operation is electrically carried out; that is, the operation of comparing the projection data with the sum data of reconstructed image, the operation of processing the difference values resulted from the comparison to obtain the correction data, and the operation of laying the correction data over the reconstructed image to correct the latter are digitally carried out.
However, the above-described method of digitally processing images suffers from difficulties that image cannot be improved in accuracy and in resolution without increasing the number of picture elements, and the image processing operation takes a long time. In addition, rotating a picture element matrix in a digital calculation is considerably difficult in the points of memory capacity and processing time, thus being unsuitable for realtime processing of image.
In other words, when, in the image reconstructing apparatus, it is required to perform the operations of adding/subtracting and rotating with respect to the reconstructed image in a digital calculation, the accuracy and the resolution of the reconstructed image cannot be improved without greatly increasing the amount of hardware, and the image processing time is considerably long.